J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.47.5.437 on 1 May 1984. Downloaded from

Journal ofNeurology, Neurosurgery, and Psychiatry 1984;47:437-442

Occasional review pathogenesis: the neural hypothesis reexamined JN BLAU From the National Hospitals for Nervous Diseases, Queen Square, and Maida Vale, Northwick Park Hospital, Harrow and City ofLondon Migraine Clinic, London, UK

SUMMARY The hypothesis that migraine is a primary neurological disturbance with secondary vascular manifestations is tested by analysing the five phases of migraine attacks and the eight groups of recognised precipitating factors. Accessory evidence from cerebral blood flow and EEG recordings taken during attacks is also considered. The evidence supports the concept that the sensory cortex and could be initiating sites for migraine attacks, and indicates that a neurological mechanism, suggested by Liveing and Gowers 100 years ago, remains viable and needs to be considered in future research. Protected by copyright.

A vascular cause for migraine has been accepted for aspect of the patient. The of the surface must several decades, on the assumption that extracranial be due to contraction of the arteries, and the flush- vasodilatation produced the characteristic throbbing ing of the skin due to their dilatation, and it is .' The extracranial hypothesis was chal- assumed that a corresponding condition of the ves- lenged because facial pallor is common during mig- sels of the is the cause of the derangement of raine episodes and because patients with temporal functions.... According to the other and alternative arteritis do not vomit or have .2 Subse- explanation of the disease, the primary derangement quent observations revealed that in 49/50 patients is of -cells of the brain. Their function from examined during migraine attacks, some arose time to time is disturbed in a peculiar manner, and from the meninges. This site would account for the the visible vaso-motor disturbance is secondary in cardinal symptoms of migraine, headache, origin". and photophobia, akin to the symptomatology of More recently clinicians supporting a neurological . cause include Appenzeller,' Bruyn,6 Johnson,7 No one questions that blood vessels are involved Pearce,8 Sjaastad,9 and Ziegler.'0 http://jnnp.bmj.com/ in the migraine process, but the hypothesis proposed here is that the vasomotor features are secondary to Hypothesis neural stimulation. That migraine is a neurogenic disorder is not novel, being favoured 100 years ago Migraine is a primary neurological disturbance with by Liveing and supported by Gowers4 who wrote: secondary vasomotor changes. "Two chief theories have been held regarding the origin of attacks. One is based upon the alteration in EXAMINATION OF HYPOTHESIS the state of the vessels that is so conspicuous in the The hypothesis is tested by analysing the five stages on September 29, 2021 by guest. of migraine attacks, the eight groups of recognised precipitants that induce attacks and evidence from and cerebral blood Address for reprint requests: Dr JN Blau, The National Hospital of (EEG) Nervous Diseases, Queen Sq, London WC1N 3B9, UK flow (CBF) studies carried out during the headache phase of migraine attacks. Received 18 November 1983. Accepted 8 December 1983 Based on a communication delivered to the American Association The five phases of migraine attacks for the Study of Headache held at Washington DC, USA, in June Prodromal symptoms The separation of prod- 1981. romes from the "I led me to favour a neurologi- 437 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.47.5.437 on 1 May 1984. Downloaded from

438 Blau cal basis for migraine having previously supported a toms would therefore require spasm sequentially of vasomotor theory.2 None of the is expl- successive branches on the posterior aspect of the icable on a vascular basis; they are all in the mind or middle cerebral artery without altering the blood arise from the brain. Prodromal symptoms occur supply anteriorly. "How the arrangement could sub- 1-24 hours before the advent of an aura or serve a slow gradual extremely orderly spread of an headache and consist of altered behaviour, mood ischaemic process staggers the imagination".'5 We changes, food craving, particularly for carbohydrate, are forced to agree with Miller Fisher that this yawning, altered bowel frequency, feeling unduly notion is untenable. tired or inappropriately cold. (3) Lashley'6 charted his own migrating teichopsiae A major differential between migraine and other more than 100 times, the slowly enlarging scintillat- is the concomitant alimentary distur- ing arc that "drifts across the " leaving bance; craving before, and vomiting during, normal vision behind. He calculated that this pro- and restricted food tolerance after the headache. cess would transgress the calcarine cortex at a rate of Hence the conclusion in a recent review'2 that "the 3 mm per minute. However, vasospasm or integration of the impulses causing appetite regula- ischaemia could not explain this classic phenomenon tion takes place mainly in the hypothalamus" is according to Miller Fisher, who suggested that only relevant. Further, altered intestinal sympathetic and a "neuro-electric" disturbance comparable with parasympathetic activity results from experimental Leao's"7 spreading could provide a poss- destruction of the hypothalamus, whose nuclei have ible explanation. , dopaminergic and glucose receptors (4) Finally, Miller Fisher found the arteriographic and even respond to iontophoretic application of evidence obtained during migraine attacks conflict- free fatty acids; each of these chemicals has been ing. implicated in migraine precipitation. The recent suc- Headache Phase Patients carefully questioned say cessful trial of a peripheral dopamine antagonist, their headache is throbbing for some ofthe time, but Protected by copyright. Motilium, in aborting migraine during the pro- it is a steady ache for the remainder of the headache dromal phase'3 is relevant, so is the efficacy of phase. But, even a throbbing pain need not be , a central dopamine antagonist, primarily vascular: a , throbbing or aching, taken early during the headache phase.'4 is not a primary vascular disorder, but a vascular Feeling cold when hungry and warm after a meal response to local . is a common physiological experience. Impaired Neurological symptoms during the headache temperature control is also closely linked with appe- phase include photophobia, , general tite in hypothalamic damaged animals.'2 It may irritability, hypersensitivity to vibration and smells, therefore be significant that many migraineurs com- lack of concentration, sleepiness, yawning, tempera- plain of feeling cold before and during a migraine: ture lability and occasionally patients mention they attempt to get warm and managing to do so increased libido in spite of pain. Clearly more than heralds or accompanies the end of an attack. extracranial vasodilatation is involved but there is The classic aura Most of us have accepted for no way of determining at present whether these many years that at least the aura of migraine must be neural disturbances during the headache phase are vascular in origin. Surprisingly Miller Fisher,'5 one primary or secondary to cerebro-vascular derange- http://jnnp.bmj.com/ of the world's authorities on cerebro-vascular dis- ments, hence further consideration of these symp- ease, has presented compelling arguments why the toms cannot help our analysis. Nevertheless, as dis- aura cannot be explained on a vascular basis: cussed later, experimentally, vascular changes fol- (1) The gradual spread of the visual or other sen- low within milliseconds of metabolic changes in the sory aura taking 10-30 minutes from beginning to .'8 end does not occur, in his experience, in cerebrovas- Sleep Resolution The fact that sleep, a neurological cular disease-a most telling point bearing in mind event, is the most common way that migraine Miller Fisher's many original contributions and his attacks resolve'9 does not necessarily mean that the on September 29, 2021 by guest. 30 years' experience in this field. underlying migraine process is neurological. How- (2) Miller Fisher's own dissection showed that as ever, looking at the attack as a whole, tiredness and the middle cerebral artery ascends the Rolandic yawning begin as prodromes before the headache fissure it sends branches anteriorly to the motor cor- starts, continue during and persist after the tex and posteriorly to the sensory cortex. Unilateral headache phase, indicating a neurological sub- paraesthesiae without motor involvement beginning stratum to the whole attack. If we add that sleep in the fingers, spreading slowly up the arm and then could have a cerebral and mental restorative func- down the leg, are well recognised as a sensory aura tion,20 and that mental stress or insufficient sleep can in migraine. The slow progression of sensory symp- induce attacks (see next section) then sleep resolu- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.47.5.437 on 1 May 1984. Downloaded from

Migraine pathogenesis: the neural hypothesis ree-xamined 439 Protected by copyright.

Prodrome Aura Headache Sleep Recovery phase phase Fig Phases and symptoms during complete migraine attacks. tion assumes greater significance. phenylethylamine probably activate blood vessels Recovery Symptoms Forty-seven out of fifty although induces EEG changes;22 furth- patients continued with symptoms for about a day ermore, it has been suggested recently that there are after the headache had disappeared:'9 mood and specific receptors for phenylethylamine in the rat mental changes, yawning and tiredness were most hypothalamus.23 Alcohol acts both on and frequent, but food intolerance and fluid imbalance blood vessels. However, the migraine that follows also featured. Because symptoms persist when the alcohol intake does not begin at the time of mental, headache has resolved and because they are explic- behavioural or vascular effects, but on waking the able on a neurological basis, these continuing symp- following morning after a night' s sleep: deep sleep is http://jnnp.bmj.com/ toms support a neurological hypothesis and enable a therefore a possible cause. Another possible new view of migraine, illustrated in the figure. mechanism is that ethanol induces hypoglycaemia If correct then the headache of migraine is like the which began in fasting subjects four hours after and roof of a house and we have been neglecting the continued for seven hours or longer after intraven- main building, let alone its foundations. ous injection.24 Sleep. Too much or too little sleep provokes Precipitating factors of individual migraine attacks headaches in 5% of the normal population and in a

Psychological Stress. Even the most ardent pro- proportion of migraineurs. Admittedly vascular on September 29, 2021 by guest. tagonist of the platelet theory of migraine, Dr Edda changes accompany sleep which nevertheless is Hanington,21 has stated that "stress is the most primarily neurologically controlled and mentally common precipitant of migraine". If, as we believe, influenced. the mind is inside the brain, then neural stimulation Hormonal variation in menstruating women. A operates in psychological stress. proportion of women have their migraine attacks Hunger or delaying meals does not affect blood ves- strictly related to their menstrual cycle (Blau and sels directly but does act on the brain as a whole: Ratcliffe, unpublished observations). witness the irritability of children waiting for a meal. is controlled by the hypothalamic-pituitary-ovarian Cheese or chocolate via tyramine or axis, and most women know from personal experi- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.47.5.437 on 1 May 1984. Downloaded from

440 Blau ence that a menstrual period can be brought on early cerebral capillaries.28 Similar considerations could or missed as a result of emotional stress. Further- apply to an increase in the total amount of CSF. more, the contraceptive pill, which can accentuate migraine, provokes depression in some, but affects Electroencephalography (EEG) blood vessels in others. EEG abnormalities in migraineurs between attacks Environmental stimuli. Environmental triggers are range from 20-55% in different series.29 More transmitted directly to the central by importantly, EEG disturbances during attacks out- the special senses: light, sound, smells, the first being last the headache by several days.30 Also prolonga- a more common precipitant than the others. Travel- tion of the P2 wave in visual evoked potentials in ling, stimulating the nervous system through the migraine subjects has been ascribed to either eyes and labyrinth, also provokes attacks. In this "ischaemic or synaptic damage due to a neurot- context, aeroplane journeys induce migraines poss- ransmitter derangement".3' The fact that EEG and ibly via mild anoxia, cabins are pressurised only to evoked potentials abnormalities can be detected 6000 ft, or via stress. between migraine attacks, when there is no Local in the head and neck can precipitate headache and no apparent vasomotor changes, lends migraine. These nociceptive stimuli are clearly support to a neurological substratum. neural. Allergy. If current claims are substantiated, then Experimental observations this stimulus is more likely to be vascularly medi- Cerebral metabolic studies. Modern dynamic ated. studies still support Roy and Sherrington' s32 Exercise produces vascular and glucose changes, but hypothesis that altered cerebral metabolism causes exercise-induced migraine usually begins after the calibre variation in cerebral vessels, and not vice game is finished, and not during vasodilation. Foot- versa: using carbon labelled deoxyglucose, ballers' migraine,25 caused by heading the ball, Sokoloff'8 showed that focal metabolic alterations Protected by copyright. directly affects the brain by jolting. Competitive occurred first and were followed by vascular reac- sport in general is a great psychic stimulus. tions within seconds; hypoxia hypoglyaemia and/or Consideration of precipitating factors show that amphetamine, each increased cerebral blood flow;33 some stimulate the nervous system directly, others physiological stimulation of the cortex with evoked affect blood vessels, and some both. potentials resulted in regional microflow increase within 1-2 seconds of stimulation.34 Direct evidence ofbrain involvement before and dur- Ledo's spreading depression and cerebral blood ing headache phase flow Although Leao's'7 experiments have been Two migraineurs each had a craniotomy for a benign criticised for being unphysiological, such as touching condition leaving a skull defect, and were studied in the cortex or using tetanising currents, he also subsequent migraine attacks. In Goltman's26 case reported his spreading cortical silence after intermit- there is published photographic evidence; Dr tent retinal stimulation in rabbits with flashing lights, Macdonald Critchley had a similar patient, a doctor highly relevant to precipitation of migraine. It is in 1928, but the observations were not published. therefore of great interest that Olesen et a135 in

Photographs from the first patient26 show an classic migraine reported variation in cerebral blood http://jnnp.bmj.com/ unduly slack skull decompression some hours before flow transgressing the cortex at a rate similar to that the attack. As the headache developed, the decom- demonstrated by Leao, concluding that their obser- pression became tense and later, when the headache vations were not compatible with cortical vascular and vomiting were at their height, bulged outwards. changes. After the attack the decompression returned to its normal slackness. SUMMARY OF ARGUMENT There is therefore no doubt that there can be The points supporting a neurological pathogenesis intracranial involvement hours before the onset of for migraine are: on September 29, 2021 by guest. headache. Brain swelling or increased cerebrospinal (1) Attacks can be precipitated by stimuli from fluid (CSF) production during the headache means inside the nervous system (stress or sleep) or direct that extractional vasodilatation is not solely respons- excitation of neural pathways from the environment ible for the symptoms. Clearly there must be a vas- (light) or pain in the head, without the intervention cular or an oedematous element but no evidence of blood vessels. whether the initial disturbance of function is of vas- (2) Neurological prodromes occur hours before the cular or neural origin. However, cerebral oedema headache or aura: yawning, tiredness or food crav- can be produced by centrally released vasopressin ings, symptoms that implicate the hypothalamus. and noradrenalin,27 by opening tight junctions in (3) The classic sensory aura, explicable only on a J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.47.5.437 on 1 May 1984. Downloaded from

Migraine pathogenesis: the neural hypothesis reexamined 441 neurological and not on a vascular basis, is not seen standing of altered neural transmission. However, in . looking at the migraine process as a whole, the (4) Neurological symptoms persist throughout the brain, and particularly the hypothalamus, would headache phase; yawning, tiredness, irritability and seem sites worthy of attention. lack of concentration. (5) Sleep resolves attacks. References (6) Neurological symptoms persist for hours after 'Wolff HG. Headache and other Head Pain. New York: the headache has resolved: mood changes, mental Oxford University Press, 1963:269. tiredness and yawning. 2 Blau JN. Migraine: a vasomotor instability of the (7) Altered cerebral metabolism causes calibre var- meningeal circulation. Lancet 1978;i: 1136-9. iation in cerebral vessels-the Roy and Sherrington 3Blau JN, Dexter SL. The site of pain origin during mig- hypothesis, is maintained by modem isotope studies. raine attacks. Cephalalgia 1981;i: 143-7. Cerebral blood flow variations during classical mig- Gowers WR. A Manual of Diseases of the Nervous Sys- raine attacks are not explicable on a vascular basis. tem. Vol II. London: Churchill, 1888:789. vascular (8) EEG changes can occur before, during and after Appenzeller 0. Hypothesis: pathogenesis of headache of the migrainous type. The role of impaired the headache phase. central inhibition. Headache 1975;15:177-9. I conclude from this analysis of the sequential 6 Bruyn GW. Putative pathomechanisms of migraine. A stages of migraine attacks and their precipitants, as critical appraisal. J Drug Res 1980;5:35-43. well as from cerebral blood flow and EEG observa- 1 Johnson ES. A basis for migraine -the auto- tions, that a neurological mechanism could be nomic theory reappraised. Postgrad Med J responsible. 1978;54:231-43. 8 Pearce J. Migraine, Clinical Features, Mechanisms and Discussion Management. Springfield, Ill: Thomas, 1969:91. 9 Sjaastad 0. Vascular and biochemical changes in mig- Protected by copyright. raine. In: Saxena PR ed. Migraine and Related What is the site of the lesion? If the hypothesis is Headaches. Rotterdam: Erasmus University, correct, we ought to attempt to answer the classical 1975:55-69. neurological question, "Which site(s) could be '° Ziegler DW, Hassanien RS, Kodanaz A, Meek JC. Cir- responsible?" Visual or auditory stimuli causing cadian rhythms of plasma cortisol in migraine. J attacks will first be transmitted to their respective Neurol Neurosurg Psychiatry 1979;42:741-8. areas of cortex. Hunger or lack of sleep are pre- " Blau JN. Migraine prodromes separated from the aura: sumed to be monitored in the hypothalamus. Stress, complete migraine. Br Med J 1980;281:658-60. a mental response, is more difficult to localise: the 12 Morley JE, Levine AS. The central control of appetite. whole brain could be involved; alternatively we Lancet 1983;i:398-401. Waelkens J. in the prevention of complete know that stress influences the hypothalamus. classical migraine. Br Med J 1982;284:944. Hence I propose that specific sensory areas of the 14 Wilkinson M, Williams K, Leyton M. Observations on cortex and the hypothalamus can act as sites where the treatment of an acute attack of migraine. Res Clin migraine attacks may be initiated. Stud Headache 1978;6:141-6. What is the nature of the lesion? This, the second 5 Fisher CM. Cerebral ischaemia-less familiar types. Clin

classical question of the neurologist, has to be mod- Neurosurg 1971;18:267-335. http://jnnp.bmj.com/ ified to, "What is the nature of the disturbances of 16 Lashley KS. Patterns of cerebral integration indicated by function?" Because anoxia, hypoglycaemia or of migraine. Arch Neurol Psychiatry alcohol can induce migraine as well as a similar 1941;46:331-9. '' Leao AAP. Spreading depression of activity in the cere- symptom-complex in non-migrainous subjects bral cortex. J Neurophysiol 1944;7:359-90. (headache, behavioural changes, hunger and 18 Sokoloff L. Local cerebral energy metabolism: its rela- nausea) I propose that we are dealing with a distur- tionship to functional activity and blood flow. Cere- bance in the oxidative pathway of the neuronal bral vascular smooth muscle and its control. Ciba metabolism, suggested by Amery,'6 and suspect that Foundation Symposium 56 (New series). Amsterdam: on September 29, 2021 by guest. a number of neurotransmitters may be involved. Elsevier, 1978:171-91. "9 Blau JN. Resolution of migraine attacks: Sleep and the Conclusion recovery phase. J Neurol Neurosurg Psychiatry 1982;45:223-6. 20 Adam K, Oswald I. Sleep is for tissue restoration. J R We ought not to throw out the vascular hypothesis Col Phys 1977;11:376-88. because blood vessels are involved, although, if the 21 Hanington E. Migraine, a blood disorder? J Drug Res hypothesis is correct, secondarily to neural stimula- 1980;5:29-34. tion. Hence studying the autonomic control of the 22 Scott DR, Moffett A, Swash M. Observations on the affected blood vessels could lead to a deeper under- relation of migraine and . An electro- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.47.5.437 on 1 May 1984. Downloaded from

442 Blau encephalographic, psychological and clinical study 30 Lapkin ML, French JH, Golden GS, Rowan AJ. The using oral tyramine. Epilepsia 1972;13:365-75. electroencephalogram in childhood basilar migraine. 23 Paul SM, Hulihan-Giblin B, Skolnik P. Amphetamine (Minneap) 1977;27:580-3. binding to rat hypothalamus: relation to anorexic 31 Kennard C, Gawel M, Rudolph M, Rose CF. Visual potency of phenylethylamines. Science evoked potentials in migraine subjects. Res Clin Stud 1982;218:487-90. Headache 1978;6:73-80. 24 Wilson NM, Brown PM, Juul SM, Prestwich SA, Sons- 32 Roy CS, Sherrington CS. On the regulation of the blood ken PH. Glucose turnover and metabolic and hor- supply of the brain. J Physiol (Lond) 1890;11:85- monal changes in ethanol-induced hypoglycaemia. Br 108. Med J 1981;282:849-53. 33 Nilsson B, Rehncrona S, Siesjo BK. Coupling of cerebral 25 Matthew WB. Footballers' migraine. Br Med J metabolism and blood flow in epiletic , 1972;2:326-7. hypoxia and hypoglycaemia. Cerebral vascular 26 Goltman AM. The mechanism of migraine. J Allergy smooth muscle and its control. Ciba Foundation Sym- 1935/36;7:351-5. posium 56 (New series). Amsterdam: Elsevier, 27 Raichle ME, Grubb RL, Eichling JO. Central neuroen- 1978:199-214. docrine regulation of brain water permeability. Cere- 34 Libbers DW, Leniger-Follert E. Capillary blood flow in bral vascular smooth muscle and its control. Ciba the brain cortex during changes in oxygen supply and Foundation Symposium 56 (New series). Amsterdam: state of activation. Cerebral vascular smooth muscle Elsevier, 1978:219-26. and its control. Ciba Foundation Symposium 56 (New 28 Rapoport SI. Osmotic opening of the blood-brain bar- series). Amsterdam: Elsevier, 1978:21-43. rier. Cerebral vascular smooth muscle and its control. Is Olesen J. Tfelt-Hansen P, Henriksen L, Larsen B. The Ciba Foundation Symposium 56 (New series). Ams- common migraine attack may not be initiated by cere- terdam: Elsevier, 1978:237-51. bral ischaemia. Lancet 1981;ii:438-40. 29 Parsonage MJ. Electroencephalographic studies in mig- 36 Amery WK. Brain hypoxia: the turning point in the raine. Pearce J, ed. Modern Topics in Migraine. Lon- genesis of migraine attacks. Cephalalgia 1982;2:83- don: 1975:72-84. Heinemann Medical, 109. Protected by copyright. http://jnnp.bmj.com/ on September 29, 2021 by guest.